Key words
ankle - modeling - kinetics - kinematics
Introduction
Achilles tendon injuries are common in runners with Achilles tendinopathy, accounting
for up to 10% of running injuries [1]
Conservative management of tendinopathies often involves exercises that seek to
gradually increase load tolerance of tendon tissue. This makes understanding
tissue-specific loading an important factor in making management decisions [2]. To this end, various strategies have been employed
to reduce or redistribute tissue loads in the lower extremity. Such strategies may
include modifying cadence, altering foot strike pattern, or using external support
systems for total body unloading. Such strategies may enable clinicians to reduce
tissue loads sufficiently to avoid pain, allow training, facilitate tissue recovery,
and gradually return to normal participation with activities [3]
[4]
[5]
[6]. Body weight
support systems and equipment offer promise in reducing loading that can include
lower body positive pressure treadmills [7] or either
fixed and movable harness systems [8]
[9]
[10]. Harness systems
appear to offer a less costly form for clinical use but reductions in loading appear
less understood.
The use of BWS during running has been previously shown to reduce vertical loading
rates, peak ground reaction forces, knee extensor moments, and patellofemoral joint
loading [4]
[11]
[12]
[13]. Reducing
mechanical loading to injured tendons with exercise therapy appears to be an
important consideration for tendon tissue reparative processes [14]
[15]. Loads that are
excessive in magnitude, duration, or frequency can interfere with adaptation and
healing and may be important to AT injury [16]. During
the course of treatment, clinicians attempt to match magnitude, volume, and rate of
loads to current tendon status and stage of healing based on a response to treatment
[6]. Therefore, AT treatment strategies may
benefit from our understanding of BWS systems as a load reduction strategy in
running.
A harness-based system that uses elastic bands to provide BWS during treadmill-based
activities has previously been shown to reduce plantar loading during walking and
running as well as ground reaction forces and patellofemoral joint loading in
running [13]. To our knowledge, the use of this device
to reduce AT loads has not been investigated. Our purpose was to determine how
harness-based BWS would influence ground reaction forces, AT-related loading
variables (AT stress, force, and impulse), foot strike angle, and cadence in
running. It was hypothesized that BWS would reduce GRF and AT-related loading
variables, while not affecting cadence and foot strike angle.
Materials and Methods
Subjects
Twenty-four healthy females (age: 25.3±2.2 yrs; mass:
63.2±5.5 kg; height: 169.7±7.2 cm) were
recruited. Power calculations to justify the sample size indicated a minimum of
18 participants would be needed to detect changes in Achilles tendon stress
based on an alpha of 0.05, correlation of 0.4, and a power of 0.9 from the peak
Achilles tendon stress differences in running between rearfoot and non-rearfoot
strike patterns based on Lyght et al. [5]. All
reported running for fitness 7–10 miles per week, had no history of
lower extremity injury or surgery for>12 months limiting running
participation, and had previous treadmill running experience. Each provided
informed consent in accordance with the following guidelines [17].
Laboratory procedures
Fifty-four markers were placed on tight-fitting clothing/shoes or the
head, trunk, pelvis, and upper and lower extremities [18]. Four markers were placed on the head, 5 on the trunk (C7 and T10
spinous processes, xiphoid process, sternal notch, right scapula), 16 on the
bilateral upper extremities (acromion, deltoid muscle, medial and lateral
humeral epicondyles, forearm, ulnar and radial styloid processes, and the second
metacarpophalangeal joint), 5 on the pelvis (both anterior superior iliac spines
and both posterior superior iliac spines, and the apex of the sacrum), and 16 on
both lower extremities (greater trochanter, anterior thigh, lateral femoral
epicondyle, anterior tibia, lateral malleolus, heel of the shoe, 2nd
and 5th metatarsophalangeal joint) [18]. All wore
the same shoe type (Zealot; Saucony, Boston, MA, USA) due to shoe-related
differences that may influence running performance characteristics. Kinematics
were collected at 180 Hz using a 12-camera motion capture system (Motion
Analysis Corp., Santa Rosa, CA, USA). Kinetics were collected at 1800 Hz using a
split belt instrumented treadmill (Treadmetrix, Park City, UT, USA). Kinematic
and analog data from the treadmill force plate were filtered with a
15 Hz low-pass Butterworth filter [19].
The BWS system (Lightspeed Lift, Duluth, MN, USA) uses an external frame
surrounding the treadmill where elastic cords are affixed to a harness worn by
each participant ([Fig. 1]). The elastic nature
of the cords coupled with their respective angle of pull provide an upward force
on the runner. A harness was placed on the legs and waist per the
manufacturer’s recommendation and a 3-minute running treadmill warm-up
was performed at 3.3 m/s (8.1 min/mile pace,
average of this study cohort). In a randomized order, participants ran under two
conditions: control (no BWS) and using an attachment point that attached the
elastic cords 30 cm above each participant’s greater trochanter
when standing from the frame to the harness. The 30-cm attachment point was
based on manufacturer guidelines in their operations manual. Each participant
ran 4 minutes for each condition, where data were obtained from the last
30 seconds. Approximately a 2-minute rest period allowed for BWS
adjustments between conditions. Before the running trials, images of the AT were
acquired using a GE LOGIQ Ultrasound P6 (Waukesha, WI, USA) with a
ML6–15 probe. Participants were positioned prone on a treatment table
with their right ankle measured to 90° in a neutral position with a
goniometer. The position was chosen to avoid the anisotropy effect by
facilitating contact between the probe and the tendon [20]. Ultrasound gel (Aquasonic Clear, Fairfield, NJ, USA) was applied
to the head of the probe. Transverse images of the AT were collected by placing
the probe 10 cm proximal to the calcaneal insertion on the posterior
aspect of the shank between the medial and lateral malleoli perpendicular to the
AT. Ultrasound AT cross-sectional areas were measured using ImageJ (Wayne
Rasbrand, National Institutes of Health, USA) software.
Fig. 1 Top left photo of the BWS system with a participant while
running on the instrumented treadmill with motion analysis markers.
During the control condition, the participant ran with the harness but
the elastic cords were not attached to the frame of the lift. This
figure provides a visual representation of the steps used to determine
the Achilles tendon (AT) loading during running.
Data analysis
Using a musculoskeletal model (Human Body Model; Motekforcelink, Amsterdam,
Netherlands) with 16 segments and 46 total degrees of freedom (DOF), muscle
forces were determined. The trunk had 3 segments: pelvis, midtrunk, and thorax,
each with 3 DOF. Shoulders had 6 DOF relative to the thorax and elbow and wrist
joints each had 2 DOF. The pelvis had 6 DOF, hip with 3 DOF, knee with 1 DOF and
subtalar and ankle joint each were modeled with 1 DOF. The knee and ankle joint
centers were 50% of the joint width and marker diameter from a static
pose. Bell et al. [21] was used to determine hip
center from pelvis markers.. The Levenberg-Marquardt algorithm was used to
minimize the musculoskeletal model pose [18] and
model kinematics were solved using a global optimization approach. Inertial
properties of body segments were based on De Leva et al. [22]. Three hundred muscle tendon units were
represented in this model where the muscle parameters, insertion points, and
wrapping points were based on Delp et al. [23].
Respective to the DOF allowed at each joint, muscle forces were estimated using
static optimization such that the sum of squared muscle forces were minimized
relative to the muscles’ maximum strengths matching the measured joint
moments from inverse dynamics [18]. To solve the
quadratic programming issue, a recurrent neural model was used [24]. Viscoelastic properties of the muscles were
not modeled in this musculoskeletal model. Muscle forces have been reported
similar to Opensim during gait performance when used with similar modeling
parameters [25]. Data were used in a custom AT
model to determine AT force [26]. Foot strike
angles were calculated as the angle between the vector connecting the marker on
the heel and the 2nd digit of foot along the anteroposterior axis in
the lab coordinate system using a custom MATLAB script for each step (MathWorks,
Inc., Natick MA, USA) based on Altman & Davis [27].
A 50 N vertical force threshold was used to determine the right leg
stance for 8 successive steps. The mean of these steps was determined for
AT-related loading variables (AT force and impulse), vertical GRF, foot strike
angle, and cadence.
Statistical analysis
A multivariate analysis with repeated measures examined for differences in peak
vGRF, peak AT stress, peak AT force, AT impulse, foot strike angle, and cadence
between control and the BWS condition in SPSS 28.0 (IBM Corporation, Armonk, NY,
USA). Alpha was set to 0.05. Effect sizes were calculated using Cohen’s
d.
Results
[Table 1] depicts the descriptive statistics,
p-values, and effect sizes of each dependent variable with and without BWS.
Multivariate analysis revealed collective differences between the control and BWS
condition (Wilk’s Lambda p<0.001). Univariate tests indicated there
were differences in peak vGRF. [Fig. 2] depicts the
ensemble averaged vGRF and AT stress for the control and BWS condition during the
stance phase of running. Peak vGRF was reduced 9.0% for the BWS condition.
Peak AT stress was reduced by 9.4% with the use of the BWS. Peak AT force
and AT impulse were reduced for the BWS condition. BWS reduced loading in these
variables 11.7 and 14.8%, respectively. Foot strike angle was not different
(p>0.05) with BWS despite increasing 2.2% while cadence decreased by
3.4% with BWS (p<0.05).
Fig. 2
a ) Ensemble average and standard deviation of the peak vertical
ground reaction force (body weight) for both conditions (control and body
weight support (BWS)) as a percentage of the stance phase of running.
b ) Ensemble average and standard deviation of the peak Achilles
tendon stress (MPa) for both conditions (control and BWS) as a percentage of
stance.
Table 1 Mean and standard deviation for vertical ground
reaction force, Achilles tendon (AT) loading variables, foot strike
angle, and cadence with and without BWS (control). P-values and effect
sizes depict differences between the BWS and control
condition.
|
Control
|
Body Weight Support
|
|
|
Variable
|
Mean (sd)
|
Mean (sd)
|
p-value, Cohen’s d
|
|
Peak Ground Reaction Force (BW)
|
2.34 (0.26)
|
2.12 (0.29)
|
p<0.0001, d=2.38
|
|
Peak Achilles Tendon Stress (MPa)
|
83.25 (12.54)
|
73.43 (14.34)
|
P<0.0001, d=0.73
|
|
Peak Achilles Tendon Force (BW)
|
7.51 (1.13)
|
6.63 (1.30)
|
p<0.0001, d=1.52
|
|
Achilles Tendon Impulse (BW*s)
|
0.82 (0.16)
|
0.70 (0.17)
|
p<0.0001, d=1.94
|
|
Foot Strike Angle (°)
|
26.65 (4.35)
|
27.25 (4.51)
|
p=0.133, d=-0.23
|
|
Cadence (Steps/minute)
|
169.13 (8.36)
|
163.63 (9.22)
|
p<.0001, d=1.91
|
Discussion
Our investigation examined how BWS would influence ground reaction forces, AT-related
loading variables (AT stress, force, and impulse), foot strike angle, and cadence in
running. All variables were reduced with the use of BWS except for foot strike
angle, which exhibited no change with BWS.
The measures of AT stress and force were higher in this study than those in previous
studies. This cohort demonstrated a peak AT stress of 83.25 MPa in the
control condition compared to previous works reporting a peak AT stress of 56.9 and
69.9 MPa [5]
[26]. Similarly, peak AT force was higher in the present study than in
previous works. This cohort of runners demonstrated peak AT force of 7.51 BW while
previous work reported peak AT force in rearfoot strike running at similar speeds of
6.46 and 5.6 BW [5]
[28].
A potential explanation for the higher values in the present study is that data
collection occurred on a treadmill compared to the other studies all being completed
during overground running. When comparing treadmill running to overground running, a
16.6% increase in plantarflexion moment along with an increase of
29.75% power absorption at the ankle has been reported during treadmill
running [29].
Peak AT stress, force, and impulse were all reduced with the BWS. Previous work has
shown that increasing cadence may also reduce AT loading regardless of foot strike
pattern. For example, Lyght et al., [5] reported that
increasing cadence by 5% from a preferred cadence reduced AT stress by 4.2
and 2.9% in rearfoot and forefoot strike conditions, respectively.
Considering that the use of BWS achieved a 9.4% reduction in AT stress, it
seems that BWS may lead to greater reductions in AT stress than altering
cadence.
Similar to changes seen with AT stress, BWS also resulted in relatively large
reductions in peak AT force. Previous work has demonstrated that a 5%
increase in cadence resulted in a 3.6 and 2.7% reduction in AT force for
rearfoot strikers and forefoot strikers, respectively. The present study
demonstrated an 11.7% decrease in peak AT force in rearfoot strikers. Thus,
in patients with symptomatic Achilles tendinopathy who do not achieve symptom
reduction with cadence manipulation, BWS may provide a reasonable therapeutic
alternative to reducing AT load in running.
There was also a large reduction in peak AT impulse using the BWS. This reduction in
impulse influences the cumulative load on the AT during stance and therefore may
have larger implications on the magnitude of cumulative loading on the AT over the
course of a training run. For example, in the control condition over the course of
running one kilometer, runners would impose a cumulative load of 693.6
BW·s/km to the Achilles tendon while running under BWS conditions
would only impose a cumulative load of 571.1 BW·s/km. This amounts
to a 17.7% reduction in the cumulative load per kilometer. Thus, the use of
BWS may provide a means to reduce the cumulative AT load in those runners that
utilize a rearfoot strike pattern.
Although our study did not directly manipulate cadence, we found that using BWS
decreased cadence. The pattern of increased BWS with decreased cadence has been
observed in other studies. Masumoto et al. [30]
reported that increasing BWS by 50 and 80% decreased preferred stride
frequency by 10.5 and 14.6%, respectively. However, AT loading variables
were not examined within their investigation. Because previous work has demonstrated
that decreasing cadence is accompanied by increased AT stress and force [5], it is reasonable to conclude that the reduction in
AT loading variables observed in this study are the result of BWS and not due to
changes in cadence.
The peak vGRF with BWS was reduced by 9.0% compared to the control. Grabowski
and Kram [30] also concluded that using a lower body
positive pressure treadmill reduced peak impact GRF linearly while running at three
different velocities (3.0, 4.0, and 5.0 m/s). It is likely that many
of the changes in AT loading are driven by the overall reduction of impacts during
running.
It should be noted that running speed has been shown to reduce AT loading variables
and that simply reducing speed from 3.9 m/s to
3.3 m/s that peak AT force and cumulative AT
loading/kilometer reduced 2.67 and 3.4%, respectively [28]. However, often the intent with interventions such
as cadence manipulation or BWS systems is to allow runners to continue to train at
an intensity that would allow for continued cardiovascular training to occur.
Although reducing running speed would lead to less metabolic demand, the use of BWS
systems to unload runners has also been shown to reduce metabolic demand at a given
running speed [31]. Implementation of BWS may need to
consider the tradeoff between metabolic demands and tissue loading when implementing
a rehabilitation strategy for AT-related injury.
AT-related injuries are common in runners due to the large forces and repetitive
loads applied to the lower extremities [32]. Running
with a rearfoot strike pattern has the potential for decreasing the risk of injury
due to the smaller loads placed on the AT [3]. As no
concurrent changes were seen in foot strike angle with BWS, this may indicate that
runners may be able to maintain their typical foot strike pattern while still
benefiting from reduced loading to the AT when using this system. This may be useful
for rehabilitation and offloading while using a typical running pattern.
BWS may also be beneficial for orthopedic conditions such as AT injuries or
rehabilitation after rupture. Saxena and Granot [33]
examined the effectiveness of using a positive pressure treadmill to obtain body
weight support in the return to activity following AT injury. The BWS group were
able to return to running nearly two weeks faster than the control group. The
progressions of dosing BWS for running included 70% BW at 13.9 weeks,
85% at 17.6 weeks, and 100% BW at 18.1 weeks. This study
demonstrated that BWS can be used for walking and running but also for neuromuscular
reeducation and during concentric strengthening exercises [33]. The cost of anti-gravity treadmills is quite expensive compared to
harness-based systems, however the precise magnitude of offloading may be less
precise [13]. At present, the amount of load reduction
dosage for various running-related injuries is largely unknown despite offloading
systems potential in being useful to rehabilitation. Nonetheless, harness-based
systems appear to offer a means of reducing tissue-specific loads, which may enhance
rehabilitation efforts.
Limitations
There are several limitations of this study. First, there is not a precise
quantification of amount of BW offloaded based on a set height of the harness
used. The elastic nature of the bungee cord system with individuals of varying
body weight and vertical oscillation influences the magnitude of offloading. The
correlation between body weight and maximum ground reaction force in this
investigation was r=0.878, indicating that 76.7% of the variance
in body weight explained the variation in peak vertical ground reaction force.
Additionally, plastic deformity of the bungee cords over time may have also
contributed to subtle changes in loading for participants tested earlier
compared to later. Thus, precise quantification of percentages of body weight
may not be possible with this system, but it appears effective in unloading
tissues of the lower extremity. Next, musculoskeletal modeling approaches have a
variety of anatomical and mechanical assumptions. However, many of these
modeling assumptions may not have influenced study findings as a repeated
measures design was used for this experiment. This study investigated only
rearfoot strike runners. How the AT is unloaded in non-rearfoot strike runners
is unknown. Lastly, changes in loading rates were not quantified in this study,
but it should be acknowledged that loading rates are an important consideration
in load management programs.
Conclusion
Overall, a harness-based BWS system was effective in reducing vGRF, AT loading
variables, and cadence while not inducing changes in foot strike pattern compared to
no body support during running. Use of this harness-based system may be useful in
rehabilitation when reduction in AT loading is desired.
